This paper will study how to estimate real time queue length at signalized intersections using sampled travel times from mobile traffic sensors. The authors first introduce the concept of Queue Fully Discharge Time (QFDT) which is related to queue lengths. The authors then show how sampled travel times, which can be directly available from mobile sensors, can be processed to generate vehicle queuing delays. Based on the observation that the queuing delay reduces linearly within a cycle, the delay pattern can beestimated by linear fitting using sampled queuing delays. QFDT can be calculated from the delay pattern, and finally be used to estimate the maximum queue length. The real-time queue length curve can then be constructed based on the maximum queue length. The model and algorithm are tested usingmicro-simulation and field data.
Abstract